notebooks/keras-model-tfidf-svd.ipynb

{"metadata":{"accelerator":"GPU","colab":{"provenance":[],"collapsed_sections":["iBcWgXT_Nr84"]},"gpuClass":"standard","kernelspec":{"name":"python3","display_name":"Python 3","language":"python"},"language_info":{"name":"python","version":"3.7.12","mimetype":"text/x-python","codemirror_mode":{"name":"ipython","version":3},"pygments_lexer":"ipython3","nbconvert_exporter":"python","file_extension":".py"}},"nbformat_minor":4,"nbformat":4,"cells":[{"cell_type":"code","source":"!lscpu","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"jJxWqE6kJXM_","outputId":"e5c387d8-ea23-448e-f675-4a0adfc5c536","executionInfo":{"status":"ok","timestamp":1671440675688,"user_tz":-420,"elapsed":572,"user":{"displayName":"Linh Nguyễn","userId":"12169619127854416765"}},"execution":{"iopub.status.busy":"2022-12-29T06:21:21.066884Z","iopub.execute_input":"2022-12-29T06:21:21.067642Z","iopub.status.idle":"2022-12-29T06:21:22.026590Z","shell.execute_reply.started":"2022-12-29T06:21:21.067511Z","shell.execute_reply":"2022-12-29T06:21:22.025465Z"},"trusted":true},"execution_count":1,"outputs":[{"name":"stdout","text":"Architecture:                    x86_64\nCPU op-mode(s):                  32-bit, 64-bit\nByte Order:                      Little Endian\nAddress sizes:                   46 bits physical, 48 bits virtual\nCPU(s):                          2\nOn-line CPU(s) list:             0,1\nThread(s) per core:              2\nCore(s) per socket:              1\nSocket(s):                       1\nNUMA node(s):                    1\nVendor ID:                       GenuineIntel\nCPU family:                      6\nModel:                           85\nModel name:                      Intel(R) Xeon(R) CPU @ 2.00GHz\nStepping:                        3\nCPU MHz:                         2000.178\nBogoMIPS:                        4000.35\nHypervisor vendor:               KVM\nVirtualization type:             full\nL1d cache:                       32 KiB\nL1i cache:                       32 KiB\nL2 cache:                        1 MiB\nL3 cache:                        38.5 MiB\nNUMA node0 CPU(s):               0,1\nVulnerability Itlb multihit:     Not affected\nVulnerability L1tf:              Mitigation; PTE Inversion\nVulnerability Mds:               Mitigation; Clear CPU buffers; SMT Host state u\n                                 nknown\nVulnerability Meltdown:          Mitigation; PTI\nVulnerability Mmio stale data:   Vulnerable: Clear CPU buffers attempted, no mic\n                                 rocode; SMT Host state unknown\nVulnerability Retbleed:          Mitigation; IBRS\nVulnerability Spec store bypass: Mitigation; Speculative Store Bypass disabled v\n                                 ia prctl and seccomp\nVulnerability Spectre v1:        Mitigation; usercopy/swapgs barriers and __user\n                                  pointer sanitization\nVulnerability Spectre v2:        Mitigation; IBRS, IBPB conditional, RSB filling\n                                 , PBRSB-eIBRS Not affected\nVulnerability Srbds:             Not affected\nVulnerability Tsx async abort:   Mitigation; Clear CPU buffers; SMT Host state u\n                                 nknown\nFlags:                           fpu vme de pse tsc msr pae mce cx8 apic sep mtr\n                                 r pge mca cmov pat pse36 clflush mmx fxsr sse s\n                                 se2 ss ht syscall nx pdpe1gb rdtscp lm constant\n                                 _tsc rep_good nopl xtopology nonstop_tsc cpuid \n                                 tsc_known_freq pni pclmulqdq ssse3 fma cx16 pci\n                                 d sse4_1 sse4_2 x2apic movbe popcnt aes xsave a\n                                 vx f16c rdrand hypervisor lahf_lm abm 3dnowpref\n                                 etch invpcid_single pti ssbd ibrs ibpb stibp fs\n                                 gsbase tsc_adjust bmi1 hle avx2 smep bmi2 erms \n                                 invpcid rtm mpx avx512f avx512dq rdseed adx sma\n                                 p clflushopt clwb avx512cd avx512bw avx512vl xs\n                                 aveopt xsavec xgetbv1 xsaves arat md_clear arch\n                                 _capabilities\n","output_type":"stream"}]},{"cell_type":"code","source":"!nvidia-smi\n","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"uWTnNtwNJe_g","outputId":"cad4c3e9-c49b-40f4-b99e-0be1faa0452a","executionInfo":{"status":"ok","timestamp":1671440677344,"user_tz":-420,"elapsed":764,"user":{"displayName":"Linh Nguyễn","userId":"12169619127854416765"}},"execution":{"iopub.status.busy":"2022-12-29T06:21:22.029239Z","iopub.execute_input":"2022-12-29T06:21:22.029701Z","iopub.status.idle":"2022-12-29T06:21:23.046201Z","shell.execute_reply.started":"2022-12-29T06:21:22.029653Z","shell.execute_reply":"2022-12-29T06:21:23.045072Z"},"trusted":true},"execution_count":2,"outputs":[{"name":"stdout","text":"Thu Dec 29 06:21:22 2022       \n+-----------------------------------------------------------------------------+\n| NVIDIA-SMI 470.82.01    Driver Version: 470.82.01    CUDA Version: 11.4     |\n|-------------------------------+----------------------+----------------------+\n| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |\n| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |\n|                               |                      |               MIG M. |\n|===============================+======================+======================|\n|   0  Tesla P100-PCIE...  Off  | 00000000:00:04.0 Off |                    0 |\n| N/A   33C    P0    26W / 250W |      0MiB / 16280MiB |      0%      Default |\n|                               |                      |                  N/A |\n+-------------------------------+----------------------+----------------------+\n                                                                               \n+-----------------------------------------------------------------------------+\n| Processes:                                                                  |\n|  GPU   GI   CI        PID   Type   Process name                  GPU Memory |\n|        ID   ID                                                   Usage      |\n|=============================================================================|\n|  No running processes found                                                 |\n+-----------------------------------------------------------------------------+\n","output_type":"stream"}]},{"cell_type":"code","source":"!python -V\n!cmake --version","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"PqfQXxslJ9jN","outputId":"2f8ae2e9-9c23-4535-96f3-1c8ee2a85d92","executionInfo":{"status":"ok","timestamp":1671440677880,"user_tz":-420,"elapsed":554,"user":{"displayName":"Linh Nguyễn","userId":"12169619127854416765"}},"execution":{"iopub.status.busy":"2022-12-29T06:21:23.048359Z","iopub.execute_input":"2022-12-29T06:21:23.048766Z","iopub.status.idle":"2022-12-29T06:21:25.095239Z","shell.execute_reply.started":"2022-12-29T06:21:23.048724Z","shell.execute_reply":"2022-12-29T06:21:25.094158Z"},"trusted":true},"execution_count":3,"outputs":[{"name":"stdout","text":"Python 3.7.12\ncmake version 3.22.5\n\nCMake suite maintained and supported by Kitware (kitware.com/cmake).\n","output_type":"stream"}]},{"cell_type":"code","source":"!pip install -q pyvi\n!pip install -q visen\n!pip install -q emoji\n!pip install -q pretty-confusion-matrix\n!pip install -q tf2onnx\n!pip install -q onnx\n!pip install onnxruntime","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"KRpUyArTJ_4I","outputId":"1c5e855e-6a85-4bf2-80a8-986ade98a63f","executionInfo":{"status":"ok","timestamp":1671440744070,"user_tz":-420,"elapsed":66195,"user":{"displayName":"Linh Nguyễn","userId":"12169619127854416765"}},"execution":{"iopub.status.busy":"2022-12-29T08:00:05.486801Z","iopub.execute_input":"2022-12-29T08:00:05.487173Z","iopub.status.idle":"2022-12-29T08:01:13.070779Z","shell.execute_reply.started":"2022-12-29T08:00:05.487139Z","shell.execute_reply":"2022-12-29T08:01:13.069570Z"},"trusted":true},"execution_count":95,"outputs":[{"name":"stdout","text":"\u001b[33mWARNING: Running pip as the 'root' user can result in broken permissions and conflicting behaviour with the system package manager. It is recommended to use a virtual environment instead: https://pip.pypa.io/warnings/venv\u001b[0m\u001b[33m\n\u001b[0m\u001b[33mWARNING: Running pip as the 'root' user can result in broken permissions and conflicting behaviour with the system package manager. It is recommended to use a virtual environment instead: https://pip.pypa.io/warnings/venv\u001b[0m\u001b[33m\n\u001b[0m\u001b[33mWARNING: Running pip as the 'root' user can result in broken permissions and conflicting behaviour with the system package manager. It is recommended to use a virtual environment instead: https://pip.pypa.io/warnings/venv\u001b[0m\u001b[33m\n\u001b[0m\u001b[33mWARNING: Running pip as the 'root' user can result in broken permissions and conflicting behaviour with the system package manager. It is recommended to use a virtual environment instead: https://pip.pypa.io/warnings/venv\u001b[0m\u001b[33m\n\u001b[0m\u001b[33mWARNING: Running pip as the 'root' user can result in broken permissions and conflicting behaviour with the system package manager. It is recommended to use a virtual environment instead: https://pip.pypa.io/warnings/venv\u001b[0m\u001b[33m\n\u001b[0m\u001b[33mWARNING: Running pip as the 'root' user can result in broken permissions and conflicting behaviour with the system package manager. It is recommended to use a virtual environment instead: https://pip.pypa.io/warnings/venv\u001b[0m\u001b[33m\n\u001b[0mCollecting onnxruntime\n  Downloading onnxruntime-1.13.1-cp37-cp37m-manylinux_2_27_x86_64.whl (4.5 MB)\n\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m4.5/4.5 MB\u001b[0m \u001b[31m11.3 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m00:01\u001b[0m0:01\u001b[0m\n\u001b[?25hRequirement already satisfied: sympy in /opt/conda/lib/python3.7/site-packages (from onnxruntime) (1.10.1)\nRequirement already satisfied: protobuf in /opt/conda/lib/python3.7/site-packages (from onnxruntime) (3.19.4)\nCollecting coloredlogs\n  Downloading coloredlogs-15.0.1-py2.py3-none-any.whl (46 kB)\n\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m46.0/46.0 kB\u001b[0m \u001b[31m4.4 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n\u001b[?25hRequirement already satisfied: packaging in /opt/conda/lib/python3.7/site-packages (from onnxruntime) (21.3)\nRequirement already satisfied: numpy>=1.21.6 in /opt/conda/lib/python3.7/site-packages (from onnxruntime) (1.21.6)\nRequirement already satisfied: flatbuffers in /opt/conda/lib/python3.7/site-packages (from onnxruntime) (1.12)\nCollecting humanfriendly>=9.1\n  Downloading humanfriendly-10.0-py2.py3-none-any.whl (86 kB)\n\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m86.8/86.8 kB\u001b[0m \u001b[31m9.0 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n\u001b[?25hRequirement already satisfied: pyparsing!=3.0.5,>=2.0.2 in /opt/conda/lib/python3.7/site-packages (from packaging->onnxruntime) (3.0.9)\nRequirement already satisfied: mpmath>=0.19 in /opt/conda/lib/python3.7/site-packages (from sympy->onnxruntime) (1.2.1)\nInstalling collected packages: humanfriendly, coloredlogs, onnxruntime\nSuccessfully installed coloredlogs-15.0.1 humanfriendly-10.0 onnxruntime-1.13.1\n\u001b[33mWARNING: Running pip as the 'root' user can result in broken permissions and conflicting behaviour with the system package manager. It is recommended to use a virtual environment instead: https://pip.pypa.io/warnings/venv\u001b[0m\u001b[33m\n\u001b[0m","output_type":"stream"}]},{"cell_type":"code","source":"!pip install --upgrade scikit-learn","metadata":{"execution":{"iopub.status.busy":"2022-12-29T09:17:04.935466Z","iopub.execute_input":"2022-12-29T09:17:04.935866Z","iopub.status.idle":"2022-12-29T09:17:15.097949Z","shell.execute_reply.started":"2022-12-29T09:17:04.935829Z","shell.execute_reply":"2022-12-29T09:17:15.096747Z"},"trusted":true},"execution_count":192,"outputs":[{"name":"stdout","text":"Requirement already satisfied: scikit-learn in /opt/conda/lib/python3.7/site-packages (1.0.2)\nRequirement already satisfied: threadpoolctl>=2.0.0 in /opt/conda/lib/python3.7/site-packages (from scikit-learn) (3.1.0)\nRequirement already satisfied: joblib>=0.11 in /opt/conda/lib/python3.7/site-packages (from scikit-learn) (1.0.1)\nRequirement already satisfied: numpy>=1.14.6 in /opt/conda/lib/python3.7/site-packages (from scikit-learn) (1.21.6)\nRequirement already satisfied: scipy>=1.1.0 in /opt/conda/lib/python3.7/site-packages (from scikit-learn) (1.7.3)\n\u001b[33mWARNING: Running pip as the 'root' user can result in broken permissions and conflicting behaviour with the system package manager. It is recommended to use a virtual environment instead: https://pip.pypa.io/warnings/venv\u001b[0m\u001b[33m\n\u001b[0m","output_type":"stream"}]},{"cell_type":"code","source":"import re\nfrom unicodedata import normalize\nfrom string import punctuation\nimport visen\nimport emoji\n\nfrom sklearn.model_selection import train_test_split\nfrom sklearn.metrics import classification_report, confusion_matrix, log_loss\nfrom sklearn.model_selection import GridSearchCV\nimport lightgbm as lgb\nimport pandas as pd\nimport pickle\nimport numpy as np\nimport matplotlib.pyplot as plt\nfrom pretty_confusion_matrix import pp_matrix\nfrom sklearn.decomposition import TruncatedSVD\nfrom sklearn.feature_extraction.text import TfidfVectorizer\n","metadata":{"id":"iXEjejimKKjn","executionInfo":{"status":"ok","timestamp":1671440895861,"user_tz":-420,"elapsed":2139,"user":{"displayName":"Linh Nguyễn","userId":"12169619127854416765"}},"execution":{"iopub.status.busy":"2022-12-29T09:17:23.076707Z","iopub.execute_input":"2022-12-29T09:17:23.077328Z","iopub.status.idle":"2022-12-29T09:17:23.085041Z","shell.execute_reply.started":"2022-12-29T09:17:23.077289Z","shell.execute_reply":"2022-12-29T09:17:23.083270Z"},"trusted":true},"execution_count":194,"outputs":[]},{"cell_type":"code","source":"import sklearn\nsklearn.__version__","metadata":{"execution":{"iopub.status.busy":"2022-12-29T09:17:33.613828Z","iopub.execute_input":"2022-12-29T09:17:33.614200Z","iopub.status.idle":"2022-12-29T09:17:33.620934Z","shell.execute_reply.started":"2022-12-29T09:17:33.614167Z","shell.execute_reply":"2022-12-29T09:17:33.619782Z"},"trusted":true},"execution_count":195,"outputs":[{"execution_count":195,"output_type":"execute_result","data":{"text/plain":"'1.0.2'"},"metadata":{}}]},{"cell_type":"code","source":"opening_ls = ['[', '{', '⁅', '〈', '⎡', '⎢', '⎣', '⎧', '⎨', '⎩', '❬', '❰', '❲', '❴', '⟦', '⟨', '⟪', '⟬', '⦃', '⦇', '⦉',\n              '⦋', '⦍', '⦏', '⦑', '⦓', '⦕', '⦗', '⧼', '⸂', '⸄', '⸉', '⸌', '⸜', '⸢', '⸤', '⸦', '〈', '《', '「', '『',\n              '【', '〔', '〖', '〘', '〚', '﹛', '﹝', '［', '｛', '｢', '｣']\n\nclosing_ls = [']', '}', '⁆', '〉', '⎤', '⎥', '⎦', '⎫', '⎬', '⎭', '❭', '❱', '❳', '❵', '⟧', '⟩', '⟫', '⟭', '⦄', '⦈', '⦊',\n              '⦌', '⦎', '⦐', '⦒', '⦔', '⦖', '⦘', '⧽', '⸃', '⸅', '⸊', '⸍', '⸝', '⸣', '⸥', '⸧', '〉', '》', '」', '』',\n              '】', '〕', '〗', '〙', '〛', '﹜', '﹞', '］', '｝', '｣']\n\nopening_brackets = {key: '(' for key in opening_ls}\nclosing_brackets = {key: ')' for key in closing_ls}\n\n### constant\nPUNC = '!\\\"#$&()*+,-–−./:;=?@[\\]^_`{|}~”“`°²ˈ‐ㄧ‛∼’'  # remove <> for number_sym and unknown_sym\nNUMBER_SYM = '<NUMBER>'\nUNKNOWN_SYM = ' <UNKNUM> '  # mix number with character, ex: 6.23A\nVIETNAMESE_CHARS = 'aàảãáạăằẳẵắặâầẩẫấậbcdđeèẻẽéẹêềểễếệfghiìỉĩíịjklmnrstoòỏõóọôồổỗốộơờởỡớợpquùủũúụưừửữứựvwxyỳỷỹýỵz'\nre_email = '([a-zA-Z0-9._%+-]+@[a-zA-Z0-9-]+\\\\.[a-zA-Z]{2,4}(\\\\.?[a-zA-Z]{2,4})?)'\nre_url = '(https?:\\/\\/(?:www\\.|(?!www))[a-zA-Z0-9][a-zA-Z0-9-]+[a-zA-Z0-9]\\.[^\\s]{2,}|www\\.[a-zA-Z0-9][a-zA-Z0-9-]+[a-zA-Z0-9]\\.[^\\s]{2,}|https?:\\/\\/(?:www\\.|(?!www))[a-zA-Z0-9]+\\.[^\\s]{2,}|www\\.[a-zA-Z0-9]+\\.[^\\s]{2,})'\nre_url2 = 'http[s]?://(?:[a-zA-Z]|[0-9]|[$-_@.&+]|[!*\\\\(\\\\),]|(?:%[0-9a-fA-F][0-9a-fA-F]))+'\nre_image = '(tập[\\s_]tin|hình|file|image|imagesize).*?(jpg|svg|png|gif|jpeg|ogg|tif|width)'\nre_num_and_decimal = '[0-9]*[,.\\-]*[0-9]*[,.\\-]*[0-9]*[.,\\-]*[0-9]*[,.\\-]*[0-9]+[.,]?'\nre_unknown = '[a-z]+[\\d]+[\\w]*|[\\d]+[a-z]+[\\w]*'\nre_vnese_txt = r'[^a-z0-9A-ZàáãạảăắằẳẵặâấầẩẫậèéẹẻẽêềếểễệđìíĩỉịòóõọỏôốồổỗộơớờởỡợùúũụủưứừửữựỳỵỷỹýÀÁÃẠẢĂẮẰẲẴẶÂẤẦẨẪẬÈÉẸẺẼÊỀẾỂỄỆĐÌÍĨỈỊÒÓÕỌỎÔỐỒỔỖỘƠỚỜỞỠỢÙÚŨỤỦƯỨỪỬỮỰỲỴỶỸÝ\\s|_]'\nre_truncate_unknown = '(<UNKNUM>\\s*)+'\n\n\nspecial_punc = {'”': '\"', '': '', \"’\": \"'\", \"`\": \"'\"}\n\n\ndef replace_all(replacer: dict, txt: str) -> str:\n    for old, new in replacer.items():\n        txt = txt.replace(old, new)\n    return txt\n\n\ndef replace_num(text: str) -> str:\n    text = re.sub(re_num_and_decimal, UNKNOWN_SYM, text)\n    return text\n\n\ndef replace_unknown(text: str) -> str:\n    text = re.sub(re_unknown, UNKNOWN_SYM, text)\n    return text\n\n\ndef unicode_normalizer(text, forms: list = ['NFKC', 'NKFD', 'NFC', 'NFD']) -> str:\n    for form in forms:\n        text = normalize(form, text)\n    return text\n\n\ndef normalize_bracket(text: str) -> str:\n    text = replace_all(opening_brackets, text)\n    text = replace_all(closing_brackets, text)\n    text = re.sub(r\"[\\(\\[].*?[\\)\\]]\", \" \", text)\n    return text\n\n\ndef remove_punc(text: str) -> str:\n    r = re.compile(r'[\\s{}]+'.format(re.escape(PUNC)))\n    text = r.split(text)\n    return ' '.join(i for i in text if i)\n\n\ndef truncate_unknown(text: str) -> str:\n    text = re.sub(re_truncate_unknown, UNKNOWN_SYM, text)\n    return text\n\n\ndef norm(text: str) -> str:\n    text = str(text)\n    text = text.lower()\n    text = text.split('\\n')[0]\n    text = unicode_normalizer(text, [\"NFKC\", \"NFKD\", \"NFD\", \"NFC\"])\n    text = replace_all(special_punc, text)\n    text = normalize_bracket(text)\n    text = replace_unknown(text)\n    text = replace_num(text)\n    text = remove_punc(text)\n    text = truncate_unknown(text)\n    text = re.sub(re_vnese_txt, \" \", text)\n    text = text.strip()\n    return visen.clean_tone(text)\n\nsample_txts = [\n    '〖𝔰𝔬𝔪𝔢 𝔣𝔞𝔫𝔠𝔶 𝔭𝔯𝔬𝔡𝔲𝔠𝔱 𝔫𝔞𝔪𝔢〗',\n    '𝖘𝖔𝖒𝖊 𝖋𝖆𝖓𝖈𝖞 𝖕𝖗𝖔𝖉𝖚𝖈𝖙 𝖓𝖆𝖒𝖊',\n    '❰some product code which will overfit this sample❱ 𝓼𝓸𝓶𝓮 𝓯𝓪𝓷𝓬𝔂 𝓹𝓻𝓸𝓭𝓾𝓬𝓽 𝓷𝓪𝓶𝓮',\n    '𝓈𝑜𝓂𝑒 𝒻𝒶𝓃𝒸𝓎 𝓅𝓇𝑜𝒹𝓊𝒸𝓉 𝓃𝒶𝓂𝑒',\n    '{𝕤𝕠𝕞𝕖 𝕗𝕒𝕟𝕔𝕪 𝕡𝕣𝕠𝕕𝕦𝕔𝕥 𝕟𝕒𝕞𝕖',\n    '☯😝ｓｏｍｅ ｆａｎｃｙ ｐｒｏｄｕｃｔ ｎａｍｅ☯😝',\n    '𝐬𝐨𝐦𝐞 𝐟𝐚𝐧𝐜𝐲 𝐩𝐫𝐨𝐝𝐮𝐜𝐭 𝐧𝐚𝐦𝐞',\n    '𝘀𝗼𝗺𝗲 𝗳𝗮𝗻𝗰𝘆 𝗽𝗿𝗼𝗱𝘂𝗰𝘁 𝗻𝗮𝗺𝗲',\n    '𝘴𝘰𝘮𝘦 𝘧𝘢𝘯𝘤𝘺 𝘱𝘳𝘰𝘥𝘶𝘤𝘵 𝘯𝘢𝘮𝘦',\n    '▄▀▄▀▄▀ 𝙨𝙤𝙢𝙚 𝙛𝙖𝙣𝙘𝙮 𝙥𝙧𝙤𝙙𝙪𝙘𝙩 𝙣𝙖𝙢𝙚 ▄▀▄▀▄▀',\n    '🌌  🎀 𝚜𝚘𝚖𝚎 𝚏𝚊𝚗𝚌𝚢 𝚙𝚛𝚘𝚍𝚞𝚌𝚝 𝚗𝚊𝚖𝚎🌌  🎀 ',\n    '(っ◔◡◔)っ ♥ some fancy product name ♥',\n]\n\nfor sample in sample_txts:\n    print(f'sample \\\"{sample}\\\", after normalize: \\\"{mask(sample)}\\\"')\n","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"YCJ-iWaYKZPU","outputId":"c1f0e0ef-3832-4f11-fd57-a07eb366b467","executionInfo":{"status":"ok","timestamp":1671440895861,"user_tz":-420,"elapsed":16,"user":{"displayName":"Linh Nguyễn","userId":"12169619127854416765"}},"execution":{"iopub.status.busy":"2022-12-29T07:44:48.015781Z","iopub.execute_input":"2022-12-29T07:44:48.016154Z","iopub.status.idle":"2022-12-29T07:44:48.041603Z","shell.execute_reply.started":"2022-12-29T07:44:48.016120Z","shell.execute_reply":"2022-12-29T07:44:48.040465Z"},"trusted":true},"execution_count":72,"outputs":[{"name":"stdout","text":"sample \"〖𝔰𝔬𝔪𝔢 𝔣𝔞𝔫𝔠𝔶 𝔭𝔯𝔬𝔡𝔲𝔠𝔱 𝔫𝔞𝔪𝔢〗\", after normalize: \"\"\nsample \"𝖘𝖔𝖒𝖊 𝖋𝖆𝖓𝖈𝖞 𝖕𝖗𝖔𝖉𝖚𝖈𝖙 𝖓𝖆𝖒𝖊\", after normalize: \"some fancy product name\"\nsample \"❰some product code which will overfit this sample❱ 𝓼𝓸𝓶𝓮 𝓯𝓪𝓷𝓬𝔂 𝓹𝓻𝓸𝓭𝓾𝓬𝓽 𝓷𝓪𝓶𝓮\", after normalize: \"some fancy product name\"\nsample \"𝓈𝑜𝓂𝑒 𝒻𝒶𝓃𝒸𝓎 𝓅𝓇𝑜𝒹𝓊𝒸𝓉 𝓃𝒶𝓂𝑒\", after normalize: \"some fancy product name\"\nsample \"{𝕤𝕠𝕞𝕖 𝕗𝕒𝕟𝕔𝕪 𝕡𝕣𝕠𝕕𝕦𝕔𝕥 𝕟𝕒𝕞𝕖\", after normalize: \"some fancy product name\"\nsample \"☯😝ｓｏｍｅ ｆａｎｃｙ ｐｒｏｄｕｃｔ ｎａｍｅ☯😝\", after normalize: \"some fancy product name\"\nsample \"𝐬𝐨𝐦𝐞 𝐟𝐚𝐧𝐜𝐲 𝐩𝐫𝐨𝐝𝐮𝐜𝐭 𝐧𝐚𝐦𝐞\", after normalize: \"some fancy product name\"\nsample \"𝘀𝗼𝗺𝗲 𝗳𝗮𝗻𝗰𝘆 𝗽𝗿𝗼𝗱𝘂𝗰𝘁 𝗻𝗮𝗺𝗲\", after normalize: \"some fancy product name\"\nsample \"𝘴𝘰𝘮𝘦 𝘧𝘢𝘯𝘤𝘺 𝘱𝘳𝘰𝘥𝘶𝘤𝘵 𝘯𝘢𝘮𝘦\", after normalize: \"some fancy product name\"\nsample \"▄▀▄▀▄▀ 𝙨𝙤𝙢𝙚 𝙛𝙖𝙣𝙘𝙮 𝙥𝙧𝙤𝙙𝙪𝙘𝙩 𝙣𝙖𝙢𝙚 ▄▀▄▀▄▀\", after normalize: \"some fancy product name\"\nsample \"🌌  🎀 𝚜𝚘𝚖𝚎 𝚏𝚊𝚗𝚌𝚢 𝚙𝚛𝚘𝚍𝚞𝚌𝚝 𝚗𝚊𝚖𝚎🌌  🎀 \", after normalize: \"some fancy product name\"\nsample \"(っ◔◡◔)っ ♥ some fancy product name ♥\", after normalize: \"some fancy product name\"\n","output_type":"stream"}]},{"cell_type":"code","source":"stopwords = pd.read_csv('/kaggle/input/data-dl-final/custom_stop_words.csv', header=0, encoding='utf-8')\nstopwords = stopwords['words'].tolist()\nmasked_data = '/kaggle/input/data-dl-final/masked_data.csv'\nmasked = pd.read_csv(masked_data, encoding='utf8')\nnon_masked = pd.read_csv('/kaggle/input/data-dl-final/shopee_data.csv', header = 0, encoding='utf8', sep='\\t', on_bad_lines='skip')","metadata":{"id":"kWK2SxT6Ko6l","executionInfo":{"status":"ok","timestamp":1671440911339,"user_tz":-420,"elapsed":7969,"user":{"displayName":"Linh Nguyễn","userId":"12169619127854416765"}},"execution":{"iopub.status.busy":"2022-12-29T06:22:22.302941Z","iopub.execute_input":"2022-12-29T06:22:22.303532Z","iopub.status.idle":"2022-12-29T06:22:28.685281Z","shell.execute_reply.started":"2022-12-29T06:22:22.303496Z","shell.execute_reply":"2022-12-29T06:22:28.684260Z"},"trusted":true},"execution_count":7,"outputs":[]},{"cell_type":"code","source":"print(masked.shape)\nprint(non_masked.shape)\nmasked","metadata":{"id":"ImsucEvvKrKK","colab":{"base_uri":"https://localhost:8080/"},"outputId":"37f067c6-f8f7-4513-a11f-dfabc5206b6c","executionInfo":{"status":"ok","timestamp":1671440911358,"user_tz":-420,"elapsed":107,"user":{"displayName":"Linh Nguyễn","userId":"12169619127854416765"}},"execution":{"iopub.status.busy":"2022-12-29T06:22:28.686758Z","iopub.execute_input":"2022-12-29T06:22:28.687113Z","iopub.status.idle":"2022-12-29T06:22:28.706134Z","shell.execute_reply.started":"2022-12-29T06:22:28.687077Z","shell.execute_reply":"2022-12-29T06:22:28.705179Z"},"trusted":true},"execution_count":8,"outputs":[{"name":"stdout","text":"(1138950, 2)\n(1138950, 2)\n","output_type":"stream"},{"execution_count":8,"output_type":"execute_result","data":{"text/plain":"                                                    sample               label\n0           miếng dán bàn phím không dây logitech <UNKNUM>  Điện tử - Điện máy\n1        nắp tròn gắn bàn phím bluetooth cho điện thoại...  Điện tử - Điện máy\n2                        pin laptop dell latitude <UNKNUM>  Điện tử - Điện máy\n3        pin hp envy <UNKNUM> envy <UNKNUM> envy <UNKNU...  Điện tử - Điện máy\n4                đầu nối chuyển đổi dc <UNKNUM> cho laptop  Điện tử - Điện máy\n...                                                    ...                 ...\n1138945                            váy body lưới dài đen m          Thời trang\n1138946                        đầm hoa nổi tay bồng cổ u s          Thời trang\n1138947                bộ dior chữ thêu trước ngực trắng l          Thời trang\n1138948                      dép ny sọc đen trắng <UNKNUM>          Thời trang\n1138949               quần legging thỏ cho bé gái <UNKNUM>          Thời trang\n\n[1138950 rows x 2 columns]","text/html":"<div>\n<style scoped>\n    .dataframe tbody tr th:only-of-type {\n        vertical-align: middle;\n    }\n\n    .dataframe tbody tr th {\n        vertical-align: top;\n    }\n\n    .dataframe thead th {\n        text-align: right;\n    }\n</style>\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>sample</th>\n      <th>label</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>0</th>\n      <td>miếng dán bàn phím không dây logitech &lt;UNKNUM&gt;</td>\n      <td>Điện tử - Điện máy</td>\n    </tr>\n    <tr>\n      <th>1</th>\n      <td>nắp tròn gắn bàn phím bluetooth cho điện thoại...</td>\n      <td>Điện tử - Điện máy</td>\n    </tr>\n    <tr>\n      <th>2</th>\n      <td>pin laptop dell latitude &lt;UNKNUM&gt;</td>\n      <td>Điện tử - Điện máy</td>\n    </tr>\n    <tr>\n      <th>3</th>\n      <td>pin hp envy &lt;UNKNUM&gt; envy &lt;UNKNUM&gt; envy &lt;UNKNU...</td>\n      <td>Điện tử - Điện máy</td>\n    </tr>\n    <tr>\n      <th>4</th>\n      <td>đầu nối chuyển đổi dc &lt;UNKNUM&gt; cho laptop</td>\n      <td>Điện tử - Điện máy</td>\n    </tr>\n    <tr>\n      <th>...</th>\n      <td>...</td>\n      <td>...</td>\n    </tr>\n    <tr>\n      <th>1138945</th>\n      <td>váy body lưới dài đen m</td>\n      <td>Thời trang</td>\n    </tr>\n    <tr>\n      <th>1138946</th>\n      <td>đầm hoa nổi tay bồng cổ u s</td>\n      <td>Thời trang</td>\n    </tr>\n    <tr>\n      <th>1138947</th>\n      <td>bộ dior chữ thêu trước ngực trắng l</td>\n      <td>Thời trang</td>\n    </tr>\n    <tr>\n      <th>1138948</th>\n      <td>dép ny sọc đen trắng &lt;UNKNUM&gt;</td>\n      <td>Thời trang</td>\n    </tr>\n    <tr>\n      <th>1138949</th>\n      <td>quần legging thỏ cho bé gái &lt;UNKNUM&gt;</td>\n      <td>Thời trang</td>\n    </tr>\n  </tbody>\n</table>\n<p>1138950 rows × 2 columns</p>\n</div>"},"metadata":{}}]},{"cell_type":"code","source":"masked['sample'] = masked['sample'].str.replace('<UNKNUM>','')\nmasked['sample'] = masked['sample'].str.replace(r'\\s+', ' ', regex=True)\nmasked","metadata":{"colab":{"base_uri":"https://localhost:8080/","height":424},"id":"WuP67itgKzLU","outputId":"009599ab-2cbf-49fe-e4c7-f3c796fb0a29","executionInfo":{"status":"ok","timestamp":1671440921467,"user_tz":-420,"elapsed":10172,"user":{"displayName":"Linh Nguyễn","userId":"12169619127854416765"}},"execution":{"iopub.status.busy":"2022-12-29T06:22:28.707576Z","iopub.execute_input":"2022-12-29T06:22:28.707954Z","iopub.status.idle":"2022-12-29T06:22:35.504160Z","shell.execute_reply.started":"2022-12-29T06:22:28.707917Z","shell.execute_reply":"2022-12-29T06:22:35.503163Z"},"trusted":true},"execution_count":9,"outputs":[{"execution_count":9,"output_type":"execute_result","data":{"text/plain":"                                                    sample               label\n0                   miếng dán bàn phím không dây logitech   Điện tử - Điện máy\n1        nắp tròn gắn bàn phím bluetooth cho điện thoại...  Điện tử - Điện máy\n2                                pin laptop dell latitude   Điện tử - Điện máy\n3                              pin hp envy envy envy envy   Điện tử - Điện máy\n4                         đầu nối chuyển đổi dc cho laptop  Điện tử - Điện máy\n...                                                    ...                 ...\n1138945                            váy body lưới dài đen m          Thời trang\n1138946                        đầm hoa nổi tay bồng cổ u s          Thời trang\n1138947                bộ dior chữ thêu trước ngực trắng l          Thời trang\n1138948                              dép ny sọc đen trắng           Thời trang\n1138949                       quần legging thỏ cho bé gái           Thời trang\n\n[1138950 rows x 2 columns]","text/html":"<div>\n<style scoped>\n    .dataframe tbody tr th:only-of-type {\n        vertical-align: middle;\n    }\n\n    .dataframe tbody tr th {\n        vertical-align: top;\n    }\n\n    .dataframe thead th {\n        text-align: right;\n    }\n</style>\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>sample</th>\n      <th>label</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>0</th>\n      <td>miếng dán bàn phím không dây logitech</td>\n      <td>Điện tử - Điện máy</td>\n    </tr>\n    <tr>\n      <th>1</th>\n      <td>nắp tròn gắn bàn phím bluetooth cho điện thoại...</td>\n      <td>Điện tử - Điện máy</td>\n    </tr>\n    <tr>\n      <th>2</th>\n      <td>pin laptop dell latitude</td>\n      <td>Điện tử - Điện máy</td>\n    </tr>\n    <tr>\n      <th>3</th>\n      <td>pin hp envy envy envy envy</td>\n      <td>Điện tử - Điện máy</td>\n    </tr>\n    <tr>\n      <th>4</th>\n      <td>đầu nối chuyển đổi dc cho laptop</td>\n      <td>Điện tử - Điện máy</td>\n    </tr>\n    <tr>\n      <th>...</th>\n      <td>...</td>\n      <td>...</td>\n    </tr>\n    <tr>\n      <th>1138945</th>\n      <td>váy body lưới dài đen m</td>\n      <td>Thời trang</td>\n    </tr>\n    <tr>\n      <th>1138946</th>\n      <td>đầm hoa nổi tay bồng cổ u s</td>\n      <td>Thời trang</td>\n    </tr>\n    <tr>\n      <th>1138947</th>\n      <td>bộ dior chữ thêu trước ngực trắng l</td>\n      <td>Thời trang</td>\n    </tr>\n    <tr>\n      <th>1138948</th>\n      <td>dép ny sọc đen trắng</td>\n      <td>Thời trang</td>\n    </tr>\n    <tr>\n      <th>1138949</th>\n      <td>quần legging thỏ cho bé gái</td>\n      <td>Thời trang</td>\n    </tr>\n  </tbody>\n</table>\n<p>1138950 rows × 2 columns</p>\n</div>"},"metadata":{}}]},{"cell_type":"code","source":"masked['label'].value_counts()","metadata":{"id":"YxIemitjK2BU","colab":{"base_uri":"https://localhost:8080/"},"outputId":"99bf823b-2dda-424f-d7e3-fc0e3cf2bad3","executionInfo":{"status":"ok","timestamp":1671440921468,"user_tz":-420,"elapsed":302,"user":{"displayName":"Linh Nguyễn","userId":"12169619127854416765"}},"execution":{"iopub.status.busy":"2022-12-29T06:22:35.508297Z","iopub.execute_input":"2022-12-29T06:22:35.509027Z","iopub.status.idle":"2022-12-29T06:22:35.566107Z","shell.execute_reply.started":"2022-12-29T06:22:35.508995Z","shell.execute_reply":"2022-12-29T06:22:35.564986Z"},"trusted":true},"execution_count":10,"outputs":[{"execution_count":10,"output_type":"execute_result","data":{"text/plain":"unknown               314845\nĐiện tử - Điện máy    260100\nThời trang            196910\nMỹ phẩm               189308\nMẹ & Bé               177787\nName: label, dtype: int64"},"metadata":{}}]},{"cell_type":"code","source":"masked = masked[~(masked['label'] == 'unknown')]\nmasked['label'].value_counts()","metadata":{"id":"8_HiKzhCK5mA","colab":{"base_uri":"https://localhost:8080/","height":424},"outputId":"becc8337-fc15-456d-9718-797f58d4ac68","executionInfo":{"status":"ok","timestamp":1671440921468,"user_tz":-420,"elapsed":296,"user":{"displayName":"Linh Nguyễn","userId":"12169619127854416765"}},"execution":{"iopub.status.busy":"2022-12-29T06:22:35.568062Z","iopub.execute_input":"2022-12-29T06:22:35.568460Z","iopub.status.idle":"2022-12-29T06:22:35.714394Z","shell.execute_reply.started":"2022-12-29T06:22:35.568423Z","shell.execute_reply":"2022-12-29T06:22:35.713291Z"},"trusted":true},"execution_count":11,"outputs":[{"execution_count":11,"output_type":"execute_result","data":{"text/plain":"Điện tử - Điện máy    260100\nThời trang            196910\nMỹ phẩm               189308\nMẹ & Bé               177787\nName: label, dtype: int64"},"metadata":{}}]},{"cell_type":"code","source":"masked.dropna(inplace=True)","metadata":{"id":"OPVvRcIGM9H-","colab":{"base_uri":"https://localhost:8080/"},"outputId":"66d84ada-90fe-49af-acf1-8a350ee0fa10","executionInfo":{"status":"ok","timestamp":1671440921469,"user_tz":-420,"elapsed":295,"user":{"displayName":"Linh Nguyễn","userId":"12169619127854416765"}},"execution":{"iopub.status.busy":"2022-12-29T06:22:35.715895Z","iopub.execute_input":"2022-12-29T06:22:35.716361Z","iopub.status.idle":"2022-12-29T06:22:35.868138Z","shell.execute_reply.started":"2022-12-29T06:22:35.716323Z","shell.execute_reply":"2022-12-29T06:22:35.867082Z"},"trusted":true},"execution_count":12,"outputs":[{"name":"stderr","text":"/opt/conda/lib/python3.7/site-packages/pandas/util/_decorators.py:311: SettingWithCopyWarning: \nA value is trying to be set on a copy of a slice from a DataFrame\n\nSee the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n  return func(*args, **kwargs)\n","output_type":"stream"}]},{"cell_type":"code","source":"X= masked[\"sample\"]","metadata":{"id":"yApSjQhYumlk","executionInfo":{"status":"ok","timestamp":1671440921471,"user_tz":-420,"elapsed":286,"user":{"displayName":"Linh Nguyễn","userId":"12169619127854416765"}},"execution":{"iopub.status.busy":"2022-12-29T06:22:35.869618Z","iopub.execute_input":"2022-12-29T06:22:35.870393Z","iopub.status.idle":"2022-12-29T06:22:35.875444Z","shell.execute_reply.started":"2022-12-29T06:22:35.870351Z","shell.execute_reply":"2022-12-29T06:22:35.874428Z"},"trusted":true},"execution_count":13,"outputs":[]},{"cell_type":"code","source":"y = masked['label']","metadata":{"id":"lImmL4uSNhAH","executionInfo":{"status":"ok","timestamp":1671440921473,"user_tz":-420,"elapsed":286,"user":{"displayName":"Linh Nguyễn","userId":"12169619127854416765"}},"execution":{"iopub.status.busy":"2022-12-29T06:22:35.879832Z","iopub.execute_input":"2022-12-29T06:22:35.880691Z","iopub.status.idle":"2022-12-29T06:22:35.885550Z","shell.execute_reply.started":"2022-12-29T06:22:35.880656Z","shell.execute_reply":"2022-12-29T06:22:35.884603Z"},"trusted":true},"execution_count":14,"outputs":[]},{"cell_type":"markdown","source":"# **Tfidf**","metadata":{"id":"gT8JAgF3OvZF"}},{"cell_type":"code","source":"import numpy as np\nfrom sklearn.model_selection import train_test_split\nfrom sklearn.feature_extraction.text import TfidfVectorizer\nfrom sklearn.preprocessing import LabelEncoder\nimport keras\nfrom sklearn.feature_extraction.text import TfidfVectorizer\nfrom sklearn.multiclass import OneVsRestClassifier\nfrom sklearn.preprocessing import LabelEncoder\nfrom keras.models import Sequential\nfrom keras.layers import Dense, Dropout, Flatten\nimport tensorflow as tf\nimport keras\nfrom keras.wrappers.scikit_learn import KerasClassifier\nfrom keras.utils import np_utils","metadata":{"id":"8k7CRZ5KfOor","executionInfo":{"status":"ok","timestamp":1671440921479,"user_tz":-420,"elapsed":278,"user":{"displayName":"Linh Nguyễn","userId":"12169619127854416765"}},"execution":{"iopub.status.busy":"2022-12-29T07:52:35.107203Z","iopub.execute_input":"2022-12-29T07:52:35.107596Z","iopub.status.idle":"2022-12-29T07:52:35.114887Z","shell.execute_reply.started":"2022-12-29T07:52:35.107562Z","shell.execute_reply":"2022-12-29T07:52:35.113587Z"},"trusted":true},"execution_count":84,"outputs":[]},{"cell_type":"code","source":"le = LabelEncoder()\ny = le.fit_transform(y)\ny = np_utils.to_categorical(y)\ny","metadata":{"execution":{"iopub.status.busy":"2022-12-29T06:22:40.025279Z","iopub.execute_input":"2022-12-29T06:22:40.026076Z","iopub.status.idle":"2022-12-29T06:22:40.212434Z","shell.execute_reply.started":"2022-12-29T06:22:40.026038Z","shell.execute_reply":"2022-12-29T06:22:40.211300Z"},"trusted":true},"execution_count":16,"outputs":[{"execution_count":16,"output_type":"execute_result","data":{"text/plain":"array([[0., 0., 0., 1.],\n       [0., 0., 0., 1.],\n       [0., 0., 0., 1.],\n       ...,\n       [0., 0., 1., 0.],\n       [0., 0., 1., 0.],\n       [0., 0., 1., 0.]], dtype=float32)"},"metadata":{}}]},{"cell_type":"code","source":"X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.3, random_state=42)","metadata":{"id":"cmnbNhKY7N5B","executionInfo":{"status":"ok","timestamp":1671440921480,"user_tz":-420,"elapsed":279,"user":{"displayName":"Linh Nguyễn","userId":"12169619127854416765"}},"execution":{"iopub.status.busy":"2022-12-29T06:22:40.214164Z","iopub.execute_input":"2022-12-29T06:22:40.214566Z","iopub.status.idle":"2022-12-29T06:22:40.352846Z","shell.execute_reply.started":"2022-12-29T06:22:40.214528Z","shell.execute_reply":"2022-12-29T06:22:40.351848Z"},"trusted":true},"execution_count":17,"outputs":[]},{"cell_type":"code","source":"y_train","metadata":{"execution":{"iopub.status.busy":"2022-12-29T06:22:40.354334Z","iopub.execute_input":"2022-12-29T06:22:40.354733Z","iopub.status.idle":"2022-12-29T06:22:40.363510Z","shell.execute_reply.started":"2022-12-29T06:22:40.354697Z","shell.execute_reply":"2022-12-29T06:22:40.362447Z"},"trusted":true},"execution_count":18,"outputs":[{"execution_count":18,"output_type":"execute_result","data":{"text/plain":"array([[0., 1., 0., 0.],\n       [0., 0., 1., 0.],\n       [0., 0., 1., 0.],\n       ...,\n       [0., 0., 0., 1.],\n       [0., 0., 1., 0.],\n       [0., 0., 0., 1.]], dtype=float32)"},"metadata":{}}]},{"cell_type":"code","source":"tfidf_vect = TfidfVectorizer()\n# tfidf_vect.fit(X_train)#.ravel()) # learn vocabulary and idf from training set\nX_train =  tfidf_vect.fit_transform(X_train)#.ravel())\n# assume that we don't have test set before\nX_test =  tfidf_vect.transform(X_test)","metadata":{"id":"a99xf79DO1Gb","executionInfo":{"status":"ok","timestamp":1671440934557,"user_tz":-420,"elapsed":13356,"user":{"displayName":"Linh Nguyễn","userId":"12169619127854416765"}},"execution":{"iopub.status.busy":"2022-12-29T06:22:40.365256Z","iopub.execute_input":"2022-12-29T06:22:40.366057Z","iopub.status.idle":"2022-12-29T06:22:49.171557Z","shell.execute_reply.started":"2022-12-29T06:22:40.366018Z","shell.execute_reply":"2022-12-29T06:22:49.170550Z"},"trusted":true},"execution_count":19,"outputs":[]},{"cell_type":"code","source":"X_train.shape","metadata":{"execution":{"iopub.status.busy":"2022-12-29T06:22:49.173225Z","iopub.execute_input":"2022-12-29T06:22:49.173635Z","iopub.status.idle":"2022-12-29T06:22:49.182518Z","shell.execute_reply.started":"2022-12-29T06:22:49.173592Z","shell.execute_reply":"2022-12-29T06:22:49.181425Z"},"trusted":true},"execution_count":20,"outputs":[{"execution_count":20,"output_type":"execute_result","data":{"text/plain":"(576809, 67332)"},"metadata":{}}]},{"cell_type":"code","source":"X_train[0].shape[1]","metadata":{"execution":{"iopub.status.busy":"2022-12-29T06:22:49.183991Z","iopub.execute_input":"2022-12-29T06:22:49.184983Z","iopub.status.idle":"2022-12-29T06:22:49.221049Z","shell.execute_reply.started":"2022-12-29T06:22:49.184945Z","shell.execute_reply":"2022-12-29T06:22:49.220087Z"},"trusted":true},"execution_count":21,"outputs":[{"execution_count":21,"output_type":"execute_result","data":{"text/plain":"67332"},"metadata":{}}]},{"cell_type":"code","source":"# Dimensionality reduction\n","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"YjaNT0mB7G_P","outputId":"07730072-af65-41e3-f1b8-027317e8fb75","executionInfo":{"status":"ok","timestamp":1671441005449,"user_tz":-420,"elapsed":70915,"user":{"displayName":"Linh Nguyễn","userId":"12169619127854416765"}},"execution":{"iopub.status.busy":"2022-12-29T06:22:49.222808Z","iopub.execute_input":"2022-12-29T06:22:49.223215Z","iopub.status.idle":"2022-12-29T06:22:49.229427Z","shell.execute_reply.started":"2022-12-29T06:22:49.223179Z","shell.execute_reply":"2022-12-29T06:22:49.228534Z"},"trusted":true},"execution_count":22,"outputs":[]},{"cell_type":"code","source":"svd = TruncatedSVD(n_components=400, random_state=42)\nX_train = svd.fit_transform(X_train)\nX_test=svd.transform(X_test)\nprint(f\"Total variance explained: {np.sum(svd.explained_variance_ratio_):.2f}\")\nprint(X_train.shape)\nprint(y_train)","metadata":{"execution":{"iopub.status.busy":"2022-12-29T06:22:49.231272Z","iopub.execute_input":"2022-12-29T06:22:49.231685Z","iopub.status.idle":"2022-12-29T06:25:27.531645Z","shell.execute_reply.started":"2022-12-29T06:22:49.231651Z","shell.execute_reply":"2022-12-29T06:25:27.530567Z"},"trusted":true},"execution_count":23,"outputs":[{"name":"stdout","text":"Total variance explained: 0.48\n(576809, 400)\n[[0. 1. 0. 0.]\n [0. 0. 1. 0.]\n [0. 0. 1. 0.]\n ...\n [0. 0. 0. 1.]\n [0. 0. 1. 0.]\n [0. 0. 0. 1.]]\n","output_type":"stream"}]},{"cell_type":"code","source":"model = Sequential()\nmodel.add(Flatten(input_shape=(400,)))\nmodel.add(Dense(256, input_dim=200, activation='relu'))\nmodel.add(Dropout(0.3))\nmodel.add(Dense(200, activation='relu'))\nmodel.add(Dropout(0.3))\nmodel.add(Dense(160, activation='relu'))\nmodel.add(Dropout(0.3))\nmodel.add(Dense(120, activation='relu'))\nmodel.add(Dropout(0.3))\nmodel.add(Dense(80, activation='relu'))\nmodel.add(Dropout(0.3))\nmodel.add(Dense(4, activation='softmax'))\n\nmodel.compile(loss='categorical_crossentropy', optimizer='adam', metrics=['accuracy'])\nes = keras.callbacks.EarlyStopping(monitor='val_f1_m', min_delta=0, patience=3, verbose=0, mode='auto', baseline=None, restore_best_weights=False)\n\nmodel.summary()\n","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"nSsjK-sNYN_n","executionInfo":{"status":"ok","timestamp":1671441011461,"user_tz":-420,"elapsed":3842,"user":{"displayName":"Linh Nguyễn","userId":"12169619127854416765"}},"outputId":"cc646a89-bdf9-4872-c583-829cfaf7a891","execution":{"iopub.status.busy":"2022-12-29T06:25:27.536088Z","iopub.execute_input":"2022-12-29T06:25:27.543635Z","iopub.status.idle":"2022-12-29T06:25:30.534602Z","shell.execute_reply.started":"2022-12-29T06:25:27.543605Z","shell.execute_reply":"2022-12-29T06:25:30.533090Z"},"trusted":true},"execution_count":24,"outputs":[{"name":"stdout","text":"Model: \"sequential\"\n_________________________________________________________________\nLayer (type)                 Output Shape              Param #   \n=================================================================\nflatten (Flatten)            (None, 400)               0         \n_________________________________________________________________\ndense (Dense)                (None, 256)               102656    \n_________________________________________________________________\ndropout (Dropout)            (None, 256)               0         \n_________________________________________________________________\ndense_1 (Dense)              (None, 200)               51400     \n_________________________________________________________________\ndropout_1 (Dropout)          (None, 200)               0         \n_________________________________________________________________\ndense_2 (Dense)              (None, 160)               32160     \n_________________________________________________________________\ndropout_2 (Dropout)          (None, 160)               0         \n_________________________________________________________________\ndense_3 (Dense)              (None, 120)               19320     \n_________________________________________________________________\ndropout_3 (Dropout)          (None, 120)               0         \n_________________________________________________________________\ndense_4 (Dense)              (None, 80)                9680      \n_________________________________________________________________\ndropout_4 (Dropout)          (None, 80)                0         \n_________________________________________________________________\ndense_5 (Dense)              (None, 4)                 324       \n=================================================================\nTotal params: 215,540\nTrainable params: 215,540\nNon-trainable params: 0\n_________________________________________________________________\n","output_type":"stream"},{"name":"stderr","text":"2022-12-29 06:25:27.645611: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n2022-12-29 06:25:27.649958: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n2022-12-29 06:25:27.650900: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n2022-12-29 06:25:27.652062: I tensorflow/core/platform/cpu_feature_guard.cc:142] This TensorFlow binary is optimized with oneAPI Deep Neural Network Library (oneDNN) to use the following CPU instructions in performance-critical operations:  AVX2 AVX512F FMA\nTo enable them in other operations, rebuild TensorFlow with the appropriate compiler flags.\n2022-12-29 06:25:27.652621: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n2022-12-29 06:25:27.653291: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n2022-12-29 06:25:27.654282: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n2022-12-29 06:25:30.072553: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n2022-12-29 06:25:30.073476: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n2022-12-29 06:25:30.074228: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n2022-12-29 06:25:30.074912: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1510] Created device /job:localhost/replica:0/task:0/device:GPU:0 with 15381 MB memory:  -> device: 0, name: Tesla P100-PCIE-16GB, pci bus id: 0000:00:04.0, compute capability: 6.0\n","output_type":"stream"}]},{"cell_type":"code","source":"y_train","metadata":{"execution":{"iopub.status.busy":"2022-12-29T06:25:30.535850Z","iopub.execute_input":"2022-12-29T06:25:30.537132Z","iopub.status.idle":"2022-12-29T06:25:30.545619Z","shell.execute_reply.started":"2022-12-29T06:25:30.537094Z","shell.execute_reply":"2022-12-29T06:25:30.544463Z"},"trusted":true},"execution_count":25,"outputs":[{"execution_count":25,"output_type":"execute_result","data":{"text/plain":"array([[0., 1., 0., 0.],\n       [0., 0., 1., 0.],\n       [0., 0., 1., 0.],\n       ...,\n       [0., 0., 0., 1.],\n       [0., 0., 1., 0.],\n       [0., 0., 0., 1.]], dtype=float32)"},"metadata":{}}]},{"cell_type":"code","source":"model.fit(X_train, y_train, batch_size = 1024, epochs = 20, callbacks=[es])","metadata":{"colab":{"base_uri":"https://localhost:8080/","height":727},"id":"s3PaLI9xR6rJ","executionInfo":{"status":"error","timestamp":1671441012924,"user_tz":-420,"elapsed":1492,"user":{"displayName":"Linh Nguyễn","userId":"12169619127854416765"}},"outputId":"e4c84366-48cb-45a1-8a7a-f78b9024f08b","execution":{"iopub.status.busy":"2022-12-29T06:25:30.547207Z","iopub.execute_input":"2022-12-29T06:25:30.548136Z","iopub.status.idle":"2022-12-29T06:26:15.201951Z","shell.execute_reply.started":"2022-12-29T06:25:30.548105Z","shell.execute_reply":"2022-12-29T06:26:15.201038Z"},"trusted":true},"execution_count":26,"outputs":[{"name":"stderr","text":"2022-12-29 06:25:30.902864: W tensorflow/core/framework/cpu_allocator_impl.cc:80] Allocation of 922894400 exceeds 10% of free system memory.\n2022-12-29 06:25:31.957624: W tensorflow/core/framework/cpu_allocator_impl.cc:80] Allocation of 922894400 exceeds 10% of free system memory.\n2022-12-29 06:25:32.721342: I tensorflow/compiler/mlir/mlir_graph_optimization_pass.cc:185] None of the MLIR Optimization Passes are enabled (registered 2)\n","output_type":"stream"},{"name":"stdout","text":"Epoch 1/20\n564/564 [==============================] - 4s 3ms/step - loss: 0.1799 - accuracy: 0.9313\nEpoch 2/20\n564/564 [==============================] - 2s 4ms/step - loss: 0.0726 - accuracy: 0.9774\nEpoch 3/20\n564/564 [==============================] - 2s 4ms/step - loss: 0.0622 - accuracy: 0.9804\nEpoch 4/20\n564/564 [==============================] - 2s 3ms/step - loss: 0.0567 - accuracy: 0.9821\nEpoch 5/20\n564/564 [==============================] - 2s 4ms/step - loss: 0.0524 - accuracy: 0.9832\nEpoch 6/20\n564/564 [==============================] - 2s 3ms/step - loss: 0.0488 - accuracy: 0.9844\nEpoch 7/20\n564/564 [==============================] - 2s 3ms/step - loss: 0.0475 - accuracy: 0.9847\nEpoch 8/20\n564/564 [==============================] - 2s 4ms/step - loss: 0.0444 - accuracy: 0.9856\nEpoch 9/20\n564/564 [==============================] - 2s 4ms/step - loss: 0.0431 - accuracy: 0.9860\nEpoch 10/20\n564/564 [==============================] - 2s 3ms/step - loss: 0.0417 - accuracy: 0.9864\nEpoch 11/20\n564/564 [==============================] - 2s 3ms/step - loss: 0.0401 - accuracy: 0.9869\nEpoch 12/20\n564/564 [==============================] - 2s 3ms/step - loss: 0.0394 - accuracy: 0.9871\nEpoch 13/20\n564/564 [==============================] - 2s 4ms/step - loss: 0.0381 - accuracy: 0.9876\nEpoch 14/20\n564/564 [==============================] - 2s 4ms/step - loss: 0.0370 - accuracy: 0.9878\nEpoch 15/20\n564/564 [==============================] - 2s 3ms/step - loss: 0.0356 - accuracy: 0.9882\nEpoch 16/20\n564/564 [==============================] - 2s 3ms/step - loss: 0.0351 - accuracy: 0.9884\nEpoch 17/20\n564/564 [==============================] - 2s 3ms/step - loss: 0.0341 - accuracy: 0.9886\nEpoch 18/20\n564/564 [==============================] - 2s 3ms/step - loss: 0.0339 - accuracy: 0.9886\nEpoch 19/20\n564/564 [==============================] - 2s 4ms/step - loss: 0.0329 - accuracy: 0.9890\nEpoch 20/20\n564/564 [==============================] - 2s 4ms/step - loss: 0.0324 - accuracy: 0.9891\n","output_type":"stream"},{"execution_count":26,"output_type":"execute_result","data":{"text/plain":"<keras.callbacks.History at 0x7faacb49d850>"},"metadata":{}}]},{"cell_type":"code","source":"pred = model.predict(X_test)","metadata":{"id":"twtnnTK7ZFqL","executionInfo":{"status":"aborted","timestamp":1671441012927,"user_tz":-420,"elapsed":550,"user":{"displayName":"Linh Nguyễn","userId":"12169619127854416765"}},"execution":{"iopub.status.busy":"2022-12-29T06:26:15.205470Z","iopub.execute_input":"2022-12-29T06:26:15.207668Z","iopub.status.idle":"2022-12-29T06:26:26.489854Z","shell.execute_reply.started":"2022-12-29T06:26:15.207628Z","shell.execute_reply":"2022-12-29T06:26:26.488808Z"},"trusted":true},"execution_count":27,"outputs":[]},{"cell_type":"code","source":"pred","metadata":{"execution":{"iopub.status.busy":"2022-12-29T06:26:26.495088Z","iopub.execute_input":"2022-12-29T06:26:26.495411Z","iopub.status.idle":"2022-12-29T06:26:26.504908Z","shell.execute_reply.started":"2022-12-29T06:26:26.495367Z","shell.execute_reply":"2022-12-29T06:26:26.503683Z"},"trusted":true},"execution_count":28,"outputs":[{"execution_count":28,"output_type":"execute_result","data":{"text/plain":"array([[1.4221607e-10, 1.0000000e+00, 9.0081583e-13, 1.3057433e-09],\n       [1.4977683e-08, 1.0000000e+00, 5.6461953e-11, 1.5583054e-09],\n       [1.1460472e-10, 8.0769023e-11, 8.5183645e-13, 1.0000000e+00],\n       ...,\n       [3.6366878e-14, 1.5073444e-14, 1.2494369e-18, 1.0000000e+00],\n       [1.5436259e-14, 9.6720771e-19, 1.0000000e+00, 5.6087325e-27],\n       [2.2827176e-12, 6.7503544e-16, 1.0000000e+00, 1.2138475e-21]],\n      dtype=float32)"},"metadata":{}}]},{"cell_type":"code","source":"pred = np.argmax(pred, axis=1)\npred","metadata":{"execution":{"iopub.status.busy":"2022-12-29T06:26:26.506654Z","iopub.execute_input":"2022-12-29T06:26:26.507226Z","iopub.status.idle":"2022-12-29T06:26:26.519262Z","shell.execute_reply.started":"2022-12-29T06:26:26.507189Z","shell.execute_reply":"2022-12-29T06:26:26.518365Z"},"trusted":true},"execution_count":29,"outputs":[{"execution_count":29,"output_type":"execute_result","data":{"text/plain":"array([1, 1, 3, ..., 3, 2, 2])"},"metadata":{}}]},{"cell_type":"code","source":"y_test = np.argmax(y_test, axis=1)\ny_test","metadata":{"execution":{"iopub.status.busy":"2022-12-29T06:26:26.520738Z","iopub.execute_input":"2022-12-29T06:26:26.521318Z","iopub.status.idle":"2022-12-29T06:26:26.532547Z","shell.execute_reply.started":"2022-12-29T06:26:26.521283Z","shell.execute_reply":"2022-12-29T06:26:26.531640Z"},"trusted":true},"execution_count":30,"outputs":[{"execution_count":30,"output_type":"execute_result","data":{"text/plain":"array([1, 1, 3, ..., 3, 2, 2])"},"metadata":{}}]},{"cell_type":"code","source":"from sklearn.metrics import confusion_matrix,classification_report\nprint(classification_report(y_test,pred))","metadata":{"id":"QYlNI16JZj1_","executionInfo":{"status":"aborted","timestamp":1671441013018,"user_tz":-420,"elapsed":641,"user":{"displayName":"Linh Nguyễn","userId":"12169619127854416765"}},"execution":{"iopub.status.busy":"2022-12-29T06:26:26.533951Z","iopub.execute_input":"2022-12-29T06:26:26.534295Z","iopub.status.idle":"2022-12-29T06:26:26.875527Z","shell.execute_reply.started":"2022-12-29T06:26:26.534260Z","shell.execute_reply":"2022-12-29T06:26:26.874359Z"},"trusted":true},"execution_count":31,"outputs":[{"name":"stdout","text":"              precision    recall  f1-score   support\n\n           0       0.98      0.98      0.98     53481\n           1       0.98      0.98      0.98     56708\n           2       0.99      0.99      0.99     58996\n           3       0.99      0.99      0.99     78020\n\n    accuracy                           0.99    247205\n   macro avg       0.98      0.98      0.98    247205\nweighted avg       0.99      0.99      0.99    247205\n\n","output_type":"stream"}]},{"cell_type":"code","source":"def predict(batch):\n    batch = svd.transform(tfidf_vect.transform(norm(i) for i in batch))\n    return le.inverse_transform(np.argmax(model.predict(batch), axis=1))\npredict(['iphone', 'ao dai 123'])","metadata":{"execution":{"iopub.status.busy":"2022-12-29T07:46:57.361253Z","iopub.execute_input":"2022-12-29T07:46:57.361960Z","iopub.status.idle":"2022-12-29T07:46:57.475313Z","shell.execute_reply.started":"2022-12-29T07:46:57.361922Z","shell.execute_reply":"2022-12-29T07:46:57.474374Z"},"trusted":true},"execution_count":80,"outputs":[{"name":"stdout","text":"(2, 400)\n","output_type":"stream"},{"execution_count":80,"output_type":"execute_result","data":{"text/plain":"array(['Điện tử - Điện máy', 'Thời trang'], dtype=object)"},"metadata":{}}]},{"cell_type":"code","source":"import tf2onnx\nimport onnx","metadata":{"execution":{"iopub.status.busy":"2022-12-29T07:46:25.181135Z","iopub.execute_input":"2022-12-29T07:46:25.181519Z","iopub.status.idle":"2022-12-29T07:46:25.187220Z","shell.execute_reply.started":"2022-12-29T07:46:25.181485Z","shell.execute_reply":"2022-12-29T07:46:25.185576Z"},"trusted":true},"execution_count":77,"outputs":[]},{"cell_type":"code","source":"sample = ['macbook pro']\ninput_signature = [tf.TensorSpec((1, 400),tf.float64, name=\"input\")]\ndynamic_axes={'text_embedding' : {0 : 'batch_size'}, 'log_softmax' : {}}\nonnx_model, _ = tf2onnx.convert.from_keras(model, input_signature, opset=13)\nonnx.save(onnx_model, \"keras_classifier.onnx\")","metadata":{"execution":{"iopub.status.busy":"2022-12-29T08:21:22.096153Z","iopub.execute_input":"2022-12-29T08:21:22.096558Z","iopub.status.idle":"2022-12-29T08:21:22.389856Z","shell.execute_reply.started":"2022-12-29T08:21:22.096521Z","shell.execute_reply":"2022-12-29T08:21:22.388937Z"},"trusted":true},"execution_count":139,"outputs":[{"name":"stderr","text":"2022-12-29 08:21:22.135561: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n2022-12-29 08:21:22.135975: I tensorflow/core/grappler/devices.cc:66] Number of eligible GPUs (core count >= 8, compute capability >= 0.0): 1\n2022-12-29 08:21:22.136209: I tensorflow/core/grappler/clusters/single_machine.cc:357] Starting new session\n2022-12-29 08:21:22.136707: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n2022-12-29 08:21:22.137126: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n2022-12-29 08:21:22.137509: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n2022-12-29 08:21:22.138192: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n2022-12-29 08:21:22.138615: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n2022-12-29 08:21:22.138885: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1510] Created device /job:localhost/replica:0/task:0/device:GPU:0 with 15381 MB memory:  -> device: 0, name: Tesla P100-PCIE-16GB, pci bus id: 0000:00:04.0, compute capability: 6.0\n2022-12-29 08:21:22.140213: I tensorflow/core/grappler/optimizers/meta_optimizer.cc:1137] Optimization results for grappler item: graph_to_optimize\n  function_optimizer: function_optimizer did nothing. time = 0.007ms.\n  function_optimizer: function_optimizer did nothing. time = 0.001ms.\n\n2022-12-29 08:21:22.190329: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n2022-12-29 08:21:22.190926: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n2022-12-29 08:21:22.191470: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n2022-12-29 08:21:22.191995: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n2022-12-29 08:21:22.192509: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n2022-12-29 08:21:22.192930: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1510] Created device /job:localhost/replica:0/task:0/device:GPU:0 with 15381 MB memory:  -> device: 0, name: Tesla P100-PCIE-16GB, pci bus id: 0000:00:04.0, compute capability: 6.0\n2022-12-29 08:21:22.202241: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n2022-12-29 08:21:22.202562: I tensorflow/core/grappler/devices.cc:66] Number of eligible GPUs (core count >= 8, compute capability >= 0.0): 1\n2022-12-29 08:21:22.202651: I tensorflow/core/grappler/clusters/single_machine.cc:357] Starting new session\n2022-12-29 08:21:22.203004: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n2022-12-29 08:21:22.203348: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n2022-12-29 08:21:22.203683: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n2022-12-29 08:21:22.204020: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n2022-12-29 08:21:22.204314: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n2022-12-29 08:21:22.204588: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1510] Created device /job:localhost/replica:0/task:0/device:GPU:0 with 15381 MB memory:  -> device: 0, name: Tesla P100-PCIE-16GB, pci bus id: 0000:00:04.0, compute capability: 6.0\n2022-12-29 08:21:22.213271: I tensorflow/core/grappler/optimizers/meta_optimizer.cc:1137] Optimization results for grappler item: graph_to_optimize\n  constant_folding: Graph size after: 41 nodes (-12), 52 edges (-12), time = 3.974ms.\n  function_optimizer: function_optimizer did nothing. time = 0.026ms.\n  dependency_optimizer: Graph size after: 35 nodes (-6), 34 edges (-18), time = 0.376ms.\n  constant_folding: Graph size after: 35 nodes (0), 34 edges (0), time = 1.491ms.\n  function_optimizer: function_optimizer did nothing. time = 0.028ms.\n  dependency_optimizer: Graph size after: 35 nodes (0), 34 edges (0), time = 0.272ms.\n\n","output_type":"stream"}]},{"cell_type":"code","source":"import onnxruntime as ort","metadata":{"execution":{"iopub.status.busy":"2022-12-29T08:21:25.811853Z","iopub.execute_input":"2022-12-29T08:21:25.812205Z","iopub.status.idle":"2022-12-29T08:21:25.817810Z","shell.execute_reply.started":"2022-12-29T08:21:25.812174Z","shell.execute_reply":"2022-12-29T08:21:25.816643Z"},"trusted":true},"execution_count":140,"outputs":[]},{"cell_type":"code","source":"so = ort.SessionOptions()\nso.add_session_config_entry('session.load_model_format', 'ONNX')\nso.graph_optimization_level = ort.GraphOptimizationLevel.ORT_ENABLE_ALL\nso.optimized_model_filepath = '/kaggle/working/keras_classifier.onnx'\nort_sess = ort.InferenceSession('/kaggle/working/keras_classifier.onnx', providers=['CPUExecutionProvider'], sess_options=so)","metadata":{"execution":{"iopub.status.busy":"2022-12-29T08:21:26.144006Z","iopub.execute_input":"2022-12-29T08:21:26.144724Z","iopub.status.idle":"2022-12-29T08:21:26.156794Z","shell.execute_reply.started":"2022-12-29T08:21:26.144686Z","shell.execute_reply":"2022-12-29T08:21:26.155241Z"},"trusted":true},"execution_count":141,"outputs":[{"name":"stderr","text":"2022-12-29 08:21:26.149740806 [W:onnxruntime:, inference_session.cc:1458 Initialize] Serializing optimized model with Graph Optimization level greater than ORT_ENABLE_EXTENDED and the NchwcTransformer enabled. The generated model may contain hardware specific optimizations, and should only be used in the same environment the model was optimized in.\n","output_type":"stream"}]},{"cell_type":"code","source":"print(ort_sess.get_inputs()[0].name)\nprint(ort_sess.get_outputs()[0].name)","metadata":{"execution":{"iopub.status.busy":"2022-12-29T08:06:48.249567Z","iopub.execute_input":"2022-12-29T08:06:48.250249Z","iopub.status.idle":"2022-12-29T08:06:48.256113Z","shell.execute_reply.started":"2022-12-29T08:06:48.250215Z","shell.execute_reply":"2022-12-29T08:06:48.254897Z"},"trusted":true},"execution_count":115,"outputs":[{"name":"stdout","text":"input\ndense_5\n","output_type":"stream"}]},{"cell_type":"code","source":"def predict_with_onnx(batch):\n    batch = svd.transform(tfidf_vect.transform(norm(i) for i in batch))\n    print(batch.shape)\n    print(type(batch))\n    pred = ort_sess.run(None, {'input': batch})\n    pred = np.argmax(pred[0], axis=1)\n    return le.inverse_transform([pred])","metadata":{"execution":{"iopub.status.busy":"2022-12-29T08:28:21.893718Z","iopub.execute_input":"2022-12-29T08:28:21.894189Z","iopub.status.idle":"2022-12-29T08:28:21.901745Z","shell.execute_reply.started":"2022-12-29T08:28:21.894155Z","shell.execute_reply":"2022-12-29T08:28:21.899898Z"},"trusted":true},"execution_count":165,"outputs":[]},{"cell_type":"code","source":"predict_with_onnx(['ao choang'])","metadata":{"execution":{"iopub.status.busy":"2022-12-29T08:28:22.054962Z","iopub.execute_input":"2022-12-29T08:28:22.055266Z","iopub.status.idle":"2022-12-29T08:28:22.132939Z","shell.execute_reply.started":"2022-12-29T08:28:22.055240Z","shell.execute_reply":"2022-12-29T08:28:22.131760Z"},"trusted":true},"execution_count":166,"outputs":[{"name":"stdout","text":"(1, 400)\n<class 'numpy.ndarray'>\n","output_type":"stream"},{"execution_count":166,"output_type":"execute_result","data":{"text/plain":"array(['Thời trang'], dtype=object)"},"metadata":{}}]},{"cell_type":"code","source":"import pickle\nimport lzma\nimport gzip\nfrom sklearn.pipeline import Pipeline","metadata":{"execution":{"iopub.status.busy":"2022-12-29T08:22:55.097254Z","iopub.execute_input":"2022-12-29T08:22:55.097949Z","iopub.status.idle":"2022-12-29T08:22:55.104990Z","shell.execute_reply.started":"2022-12-29T08:22:55.097911Z","shell.execute_reply":"2022-12-29T08:22:55.104064Z"},"trusted":true},"execution_count":145,"outputs":[]},{"cell_type":"code","source":"pipe = Pipeline([('vectorize', tfidf_vect), ('trucate', svd)])","metadata":{"execution":{"iopub.status.busy":"2022-12-29T08:25:03.257519Z","iopub.execute_input":"2022-12-29T08:25:03.257883Z","iopub.status.idle":"2022-12-29T08:25:03.263337Z","shell.execute_reply.started":"2022-12-29T08:25:03.257845Z","shell.execute_reply":"2022-12-29T08:25:03.262258Z"},"trusted":true},"execution_count":150,"outputs":[]},{"cell_type":"code","source":"trf = pipe.transform(['ao choang'])\npred = ort_sess.run(None, {'input': trf})\npred = np.argmax(pred[0], axis=1)\nprint(le.inverse_transform([pred]))","metadata":{"execution":{"iopub.status.busy":"2022-12-29T08:29:14.756008Z","iopub.execute_input":"2022-12-29T08:29:14.756618Z","iopub.status.idle":"2022-12-29T08:29:14.832813Z","shell.execute_reply.started":"2022-12-29T08:29:14.756581Z","shell.execute_reply":"2022-12-29T08:29:14.831774Z"},"trusted":true},"execution_count":172,"outputs":[{"name":"stdout","text":"(1, 400)\n<class 'numpy.ndarray'>\n['Thời trang']\n","output_type":"stream"}]},{"cell_type":"code","source":"def predict_with_onnx_and_skpipeline(sample):\n    sample = pipe.transform([norm(sample)])\n    pred = ort_sess.run(None, {'input': sample})\n    pred = np.argmax(pred[0], axis=1)\n    return le.inverse_transform([pred])","metadata":{"execution":{"iopub.status.busy":"2022-12-29T08:35:38.700209Z","iopub.execute_input":"2022-12-29T08:35:38.700592Z","iopub.status.idle":"2022-12-29T08:35:38.706648Z","shell.execute_reply.started":"2022-12-29T08:35:38.700559Z","shell.execute_reply":"2022-12-29T08:35:38.705687Z"},"trusted":true},"execution_count":186,"outputs":[]},{"cell_type":"code","source":"predict_with_onnx_and_skpipeline('ao choang')","metadata":{"execution":{"iopub.status.busy":"2022-12-29T08:35:40.044056Z","iopub.execute_input":"2022-12-29T08:35:40.044871Z","iopub.status.idle":"2022-12-29T08:35:40.122122Z","shell.execute_reply.started":"2022-12-29T08:35:40.044816Z","shell.execute_reply":"2022-12-29T08:35:40.121006Z"},"trusted":true},"execution_count":187,"outputs":[{"execution_count":187,"output_type":"execute_result","data":{"text/plain":"array(['Thời trang'], dtype=object)"},"metadata":{}}]},{"cell_type":"code","source":"with open('Vectorizer.pkl', 'wb') as f:\n    pickle.dump(pipe, f)","metadata":{"execution":{"iopub.status.busy":"2022-12-29T08:36:48.940097Z","iopub.execute_input":"2022-12-29T08:36:48.940486Z","iopub.status.idle":"2022-12-29T08:36:49.598283Z","shell.execute_reply.started":"2022-12-29T08:36:48.940451Z","shell.execute_reply":"2022-12-29T08:36:49.597142Z"},"trusted":true},"execution_count":188,"outputs":[]},{"cell_type":"code","source":"","metadata":{},"execution_count":null,"outputs":[]}]}